Design And Implementation Of Semantic Search Subsystem For Personal Data On Mobile Phone

With the rapid development of mobile internet and the increasing popularity of mobile phones, personal mobile phones are producing a large amount of personal data, including photos, emails, short messages, calendar entries and contacts, etc. With the growing of the personal data, end users'demand of retrieving these data is also increasing. Kaleido Photo is a personal information retrieval system on mobile phone, which could retrieve the personal data on mobile phone by using fulltext retrieval technology. Kaleido Photo mainly uses photos for retrieval entrance and supports tag sharing and expansion between associated photos. Information of photos is used for event based personal data retrieval.Semantic search subsystem is the semantic searching part of Kaleido Photo. It provides a comfortable way for end users to a launch semantic search for personal data by inputting keywords. In addition, it also provides users with the semantic organization for short messages based on a conversational approach, which is convenient for users to recall the context of the messages when browsing them. Firstly, technologies related with Kaleido Photo system are introduced, including information retrieval technology, semantic desktop technology, clustering technology, semantic query construction technology, etc. Secondly, System functional demand and environment of software and hardware are described. Thirdly, the design of system architecture is proposed based on demand analysis, and a general design of the system is presented. After that, the specific design and implementation of main modules of the semantic search subsystem is given.This thesis focuses on discussing the design and implementation of the semantic search subsystem of Kaleido Photo. It details the design and implementation of data extraction module, semantic data storage module, semantic query construction module, semantic search module and conversational short message clustering module, and focuses on discussing semantic query construction algorithm and conversational short message clustering algorithm. Finally, through carrying out unit testing and integration testing, the functions of the system are validated, and it is proven that the function and performance of the system meet the design requirements.In the end, the summary of the whole thesis, current shortcomings of the system and future work are given. Also, the summary of my work and research progress during master-degree study are described.